Have you already considered CircLigase™ ssDNA Ligase from epicentre? (I guess that  there are other companies as well)

Thanks for the reply Uwe. Indeed, I have looked at CircLigase. It is definitely the enzyme I will use to circularize my ssDNA.

My only issue is that I am starting with dsDNA in the form of PCR generated amplicon and cannot think of a way to generate ssDNA.

My two choices seem to be heat or enyzmes.

Exonucleases come to mind, but will chew up both strands. (Unless there is one out there I am unaware of)

I've been thinking about doing a melt in the presence of CircLigase, where my dsDNA is converted to ssDNA that any active enzyme may then act upon as a substrate for circularization.

Of course, the enzyme will be operating out of its optimal temp, but perhaps an excess of enzyme will overcome this.

[I realized this is all overly complicated, see my next post]

There are tricks you can play to degrade one strand or the other. They’re tedious, low-yield, and inefficient, however.

The basic idea is:

 [RG really messed up the figures, sorry. Image added to show alignment.]

You PCR Amplify a dsDNA product. The amplified product now has a 5’ phosphate (PO4) on each end.

PO4-GATTACAGATTACAGATTACAGATTACAGATTACAGATTACA--OH

OH--CTAATGTCTAATGTCTAATGTCTAATGTCTAATGTCTAATGT-PO4

You dephosphorylate (e.g. with AntPhos) the product to remove the PO4’s.

        GATTACAGATTACAGATTACAGATTACAGATTACAGATTACA--OH

OH--CTAATGTCTAATGTCTAATGTCTAATGTCTAATGTCTAATGT

You then cut the product near one end with a restriction endonuclease (e.g. the ‘right’ end by conventional labeling). This revels two new 5’ PO4’s.

         GATTACAGATTACAGATTACAGATTACAGATTACAGAT--OH   

OH—CTAATGTCTAATGTCTAATGTCTAATGTCTAATGTC-PO4     

    PO4-TACA--OH

  OH-TAATGT

You can then use Lambda Exonuclease to degrade the strands with PO4. (https://www.neb.com/tools-and-resources/selection-charts/properties-of-exonucleases-and-endonucleases)

    GATTACAGATTACAGATTACAGATTACAGATTAC--OH    

                                                                                                       OH-TAATGT

This leaves an ssDNA of interest (the longer one), and an oligonucleotide that’s not of interest. You can use purification strategies like size-exclusion, differential precipitation or silica adsorption to try to purify the longer ssDNA if necessary. Alternatively, you could plan the synthesis such that the oligonucleotide won’t interfere with the final assay.

Ozlem Arat used CirQ Ligase and Split ligation in some of her work at UNC. It may be in her dissertation. (https://cdr.lib.unc.edu/record/uuid:47f664db-2d54-4773-a881-af8de8319c0f -- I think).

If you have the time and inclination, it should also be possible to put a split filamentous phage origin of replication into a plasmid or a lambda phage (a la lambda ZAP,  sold in the nineties by Stratagene) and insert the sequence of interest between the ORI halves. Then it would be a matter of transforming the construct into an E. coli host, superinfecting with a helper phage, and rescuing the resultant circular ssDNA from the phage particles in the supernatant.

Derp. I forgot a reason this is far simpler than I made it seem.

Primers are not phosphorylated unless you order them phosphorylated or phosphorylate them yourself. You can degrade one strand from your reaction by simply ordering that the strand of interest is amplified with a non-phosphorylated primers…

        GATTACAGATTACAGATTACAGATTACAGATTACAGATTACA--OH

OH--CTAATGTCTAATGTCTAATGTCTAATGTCTAATGTCTAATGT-PO4

…then degrading the products with Lambda Exo…

    GATTACAGATTACAGATTACAGATTACAGATTACAGATTACA--OH

This leaves an ssDNA of interest, unless I've missed something...